•CD r THE PHILOSOPHY OF THE UPANISHADS PRINTED BY MORRISON AND GIBB LIMITED, FOR T. & T. CLARK, EDINBURGH. LONDON: SIMPKIN, MARSHALL, HAMILTON, KENT, AND co. LIMITED. NEW YORK : CHARLES SCRIBNER'S SONS. The Religion and Philosophy of India THE PHILOSOPHY OF THE UPANISHADS BY P A U L D E U S S E N PROFESSOR OF PHILOSOPHY IN THE UNIVERSITY OF KIEL AUTHORISED ENGLISH TRANSLATION BY REV. A. S. GEDEN, M.A. TUTOR IN OLD TESTAMENT LANGUAGES AND LITERATURE, AND CLASSICS, WESLEYAN COLLEGE, RICHMOND EDINBURGH: T. & T. CLARK, 38 GEORGE STREET 1908 io? Printed . . . 1906 Reprinted . . 100S PREFACE DR. DEUSSEN'S treatise on the Upanishads needs no formal introduction or commendation to students of Indian thought who are familiar with the German language. To others I would fain hope that the translation here presented, which appears with the author's sanction, may serve to make known a work of very marked ability and of surpassing interest. As far as my knowledge extends, there is no adequate exposition of the Upanishads available in English. The best was published by Messrs. Triibner more than a quarter of a century ago, and is in many respects out of date. As traced here by the master-hand of the author, the teaching of the ancient Indian seers presents itself in clearest light, and claims the sympathetic study of all lovers of truth. For the English rendering I am alone responsible. And where I may have failed to catch the precise meaning of the original, or adequately to represent the turn of phrase, I can only ask the indulgence of the reader. Dr. Deussen s style is not easy. And if a more capable hand than mine had been willing to essay the task of trans lation, I would gladly have resigned my office. With whatsoever care I can hardly hope entirely to have vi PREFACE escaped error. But for any indication of oversight or mistake, and any suggestion for improvement, I shall be most grateful. The work has exacted many hours that could be ill spared from a very full life. If however it conduce in any way to a better understanding of the mind and heart of India I shall be amply repaid. A. S. GEDEN. RICHMOND, December 1U05. PREFACE BY THE AUTHOR THE present work forms the second part of my General History of Philosophy. It is however complete in itself; and has for its subject the Philosophy of the Upanishads, the culminating point of the Indian doctrine of the universe. This point had been already reached in Vedic, pre-Buddhist times ; and in philosophical significance has been surpassed by none of the later developments of thought up to the present day. In particular the Sankhya system has followed out lines of thought traced for it in the Upanishads, and has emphasized realistic tendencies already found there (infra, pp. 239-255). Buddhism also, though of entirely independent origin, yet betrays its indebtedness in essential points to the teaching of the Upanishads, when its main fundamental thought (nirvdnam, the removal of suffering by the removal of trishnd) meets us expressed in other words (union with Brahman by the removal of kdma) in the passage from the Brihadaranyaka quoted below.1 The thoughts of the Vedanta therefore became for India a permanent and characteristic spiritual atmosphere, which pervades all the products of the later literature. 1 Brih. 4. 4. 6, infra p. 348. vii viii PREFACE BY THE AUTHOR To every Indian Brahman to-day the Upanishnd* are what the New Testament is to the Christian. So significant a phenomenon deserved and demanded a more comprehensive treatment than it had yet obtained. And my hope is to remove in some measure the cloud which hitherto has obscured this subject, and to exhibit order and consistency in place of the confused mass of contradictory conceptions, which alone had been supposed to exist. If the result is not a uniform and unified system, there is yet found a regular historical develop ment, the key to which is an original, abrupt and daring idealism ; and this in its further progress by a twofold concession, on the one hand to traditional beliefs, and on the other to the empirical prepossessions natural to us all, was gradually developed into that which we, adopting Western phraseology if not always in a Western sense, call pantheism, cosmogonism, theism, atheism (Sahkhya), and deism (Yoga). Chap, ix., "The Unreality of the Universe" (pp. 226-239), which by its paradoxical title attracts attention and provokes contradiction, or the final survey at the close of the book (p. 396 ff.), may well serve as a first introduction to these oriental teachings. o A remarkable and at first sight perplexing feature in this entire evolution of thought is the persistence with which the original idealism holds its ground, not annulled or set aside by the pantheistic and theistic developments that have grown out of it. On the contrary it remains a living force, the influence of which may be more or less directly traced everywhere, until it is finally abandoned by the Sankhya system. Adopted by the Vedanta it is proclaimed as the only " higher knowledge" (para vidyd), PREFACE BY THE AUTHOR ix and contrasted with all those realistic developments which together with the creation and transmigration doctrines are known as the "lower knowledge" (apard vidya), and are explained as accommodations of the written revelation to the weakness of human understanding. This accommodation theory of the later Vedantist teachers is not wholly baseless, and needs correction only in the one point that this adjustment to the empirical capacity of the intellect (which works within the relations of time, space and causality) was not intentional and conscious, but unconscious. In this shape the idea of accommodation becomes a key which is fitted to unlock the secrets not only of the doctrinal developments of the Upanishads, but of many analogous phenomena in Western philosophy. For the practice of clothing metaphysical intuitions in the forms of empirical knowledge is met with not only in India, but also in Europe from the earliest times. And for that very reason no account would have been taken of it had not Kant demonstrated the incorrectness of the whole procedure, as I hope to show in detail in the later parts of my work. P. DEUSSEN. CONTENTS THE PHILOSOPHY OF THE UPANISHAPS: THE SECOND PERIOD OF INDIAN PHILOSOPHY, Oil THE CON TINUANCE AND CLOSE OF THE TIMES OF THE BR&HMANAS INTRODUCTION TO THE PHILOSOPHY OF THE UPANISHADS PAGE I THE PLACE OF THE UPANISHADS IN THE LITERATURE OF THE VEDA . 1. The Veda and its Divisions . 2*. Brahmana, Aranyaka, Upanishad . 3. The Upanishads of the three older Vedas 4' The Upanishads of the Atharvaveda . 5'. On the Meaning of the word Upanishad II. BRIEF SUMMARY OF THE HISTORY OF THE UPANISHADS 1. The earliest Origin of the Upanishads • ^ 2. The extant Upanishads 2f 3 The Upanishads in Badarfxyaiia and S ankara ^ 4'. The most important Collections of Upanishads HI THE FUNDAMENTAL CONCEPTION OF THE UPANISHADS AND ITS oo— OU SIGNIFICANCE . 1 The Fundamental Conception of the Upanishads . 2. The Conception of the Upanishads in its relation t ^ 3. ThehiCoTception of 'the Upanishads in its relation to ^ Religion xi 2 5 7 10 1C- 38 xii CONTENTS PAOR THE SYSTEM OF THE UPANISTTADS INTRODUCTION . . . . . . . 51-53 FIRST PART: THEOLOGY, OR THE DOCTRINE OF BRAHMAN I. ON THE POSSIBILITY OF KNOWING BRAHMAN . . 54-85 1. Is the Veda the Source of the Knowledge of Brahman ? . 54 2. Preparatory Means to a Knowledge of Brahman . . 60 ' 3. Sacrifice . . . . . . .61 -4. Asceticism (tapas) . . . , . .65 5. Other Preliminary Conditions . . . .70 6. The Standpoint of Ignorance, of Knowledge, and of superior Knowledge in relation to Brahman . . 74 II. THE SEARCH FOR BRAHMAN ..... 85-99 7* 1. The Atman (Brahman) as the Unity . . .85 2. Bahlki's Attempts at Explanation . . . .87 3. S'akalya's Attempts at Explanation . . . .88 4. Six inadequate Definitions . . . . .89 5. Definitions of the Atman Vais'vunara . . .90 6. Nurada's gradual Instruction . . . . .92 7. Three dift'erentAAtmans . . . . .94 8. Five different Atmans . . . . .97 III. SYMBOLICAL REPRESENTATIONS OF BRAHMAN . . 99-125 1. Introduction and Classification . . . .99 2. Brahman as Prtina and Viiyu . . . .101 3. Other Symbols of Brahman . . . . .111 4. Attempts to interpret the symbolical Representations of Brahman . . . . . . .117 5. Interpretations of and Substitutes for Ritual Practices . 119 IV. THE ESSENTIAL BRAHMAN ..... 126-157 1. Introduction . . . . . . .126 2. Brahman as Being and not-Being, Reality and not-Reality . 1/28 3. Brahman as Consciousness, Thought (cit) . . .132 4. Brahman as Bliss (dnanda) ..... 140 5. Negative Character and Unknowableness of the essential Brahman ....... 146 V. BRAHMAN AND THE UNIVERSE .... 157-179 1. Sole Reality of Brahman ..... 157 2. Brahman as the cosmical Principle .... 159 3. Brahman as the psychical Principle . . . .166 4. Brahman as a Personal God (fovani) . . . .172 THE PHILOSOPHY OF THE UPANISHADS A. INTRODUCTION TO THE PHILOSOPHY OF THE UPANISHADS I. THE PLACE OF THE UPANISHADS IN THE LITEEATUEE OF THE VEDA 1. The Veda and its Divisions IT will be remembered that our earlier investigations led to a classification of Vedic literature into four principal parts, which correspond to the four priestly offices at the Soma sacrifice ; these are the Rig, Ya.jur, Sama, and Atharvaveda, each of which comprises a Samhita, a Brah- mana, and a Sutra. The Brahmana (in the wider sense of the term) is then further divided by the exponents of the Vedanta into three orders, which as regards their contents are for the most part closely connected with and overlap one another, viz. — Vidhi, Arthavada, and Vedanta or Upanishad. The following scheme may be helpful in retaining in the memory this primary classification of the Veda : — I. Rigveda. ^ A Samhita. , Tr. . TT a A I a. Vidhi. m YTur^k * Brilhmana- <>• Arthavida. iv! Atharvavecla. J G Sfltra' [ 2 THE PHILOSOPHY OF THE UPANISHADS A further preliminary remark is that each of the above twelve parts of the Veda has been preserved as a rule not separately, but in several often numerous forms, inasmuch as each Veda was taught in different S'dkhds (literally, "branches" of the tree of the Veda), i.e. Vcdic schools, which in their treatment of the common subject- matter varied so considerably from one another that, in course of time, distinct works were produced, the contents of which nevertheless remained practically the same. In particular, each of the three ancient Vedas (in the case of the fourth the relations are usually different) comprises not one Brahmana, but several ; and similarly there exist for each Veda not one but several Upanishads. On this subject more will be found below. 2. Brdhmana, Aranyaka, Upanishad The link between the Upanishad and the Brahmana with its very different spirit is as a rule^ not direct, but established ordinarily by means of an Aranyaka or " forest-book," to the close of which the Upanishad is attached, or in which it is included. The name is given either because (as Oldenberg supposes, Prol, p. 291), on account of its mysterious character it should be imparted to the student not in the village (grdme), but outside of it (aranye, in the jungle) (cp. the narrative, Brih. 3. 2. 13, and the names raliasyam, upanishad), or because from the very beginning it was "a Brahmana appointed for the vow of the anchorite." The contents of the Aranyakas perhaps favour rather the latter con ception, so far as they consist mainly of all kinds of explanations of the ritual and allegorical speculations therein. This is only what might be expected in the life 1 Aranyaka-vrata-rApam brdhtnanam, Sayana ; see Aufrecht, Einl. zum Ait. 7?r., p. iii., and cp. fteuraen, Upan., p. 7. BRAHMAN A ARANYAKA UPANISHAD 3 of the forest as a substitute for the actual sacrificial observances, which for the most part were no longer practicable ; and they form a natural transition to the speculations of the Upanishads, altogether emancipated as these are from the limitations of a formal cult. The connecting-link is never wanting where the written tradition of a Sakha has been handed down unbroken (as is not the case with the Kdthaka, S'vetds'vatara, Maitrdyaniya), for both the Aitareyins and KausMtaTdns of the Eigveda and the Taittiriyakas and Vdjasaneyins of the Yajurveda possess together with the Samhita their Brahmana with Aranyaka and Upanishad. Even then, if in the schools of the Samaveda the name Aranyaka is not employed, yet there also the introductions to the Upanishads l bear throughout the character of Aranyakas. This succession of ritual allegorical and philosophical texts, which is really the same in all the Sakhas, may be due partly to the order of thought adopted for the pur poses of instruction, in which the Samhita would naturally be followed immediately by the Brahmana (so far as this was generally taught, cp. Oldenberg, Pro/., p. 291); the deep mysterious meaning of the ceremonies would then be unfolded in the Aranyaka ; and finally the exposition of the Upanishads would close the period of Vedic in struction. As early, therefore, as S'vet. 6. 22 and Mund. 3. 2. 6, and thenceforward, the Upanishads bore the name Veddnta (i.e. "end of the Veda"). On the other hand it is not to be denied that the order of the texts within the canon of each Sakha corresponds generally to their historical development, and that the position of the several parts affords an indication of their earlier or later date. If, however, these two factors that determined the arrangement, namely, the tendency to a systematic classification of the material for instruction and the 1 Chandogya Upan. 1-2, Upanishadbrah. 1-3. 4 THE PHILOSOPHY OF THE UPANISHADS preservation of the order of chronological development, do actually for the most part coincide in their result, this is very simply explained on the supposition that in the course of time the general interest was transferred from the ritualistic method of treatment to the allegorical, and from that again to the philosophical. Moreover, the separation of the material is by no means strictly carried out, but in all three classes, Brahmanas, Aranyakas, and Upanishads, there are found occasionally digressions of a ritual as well as allegorical or philosophical nature. Especially noteworthy, however, and demanding explana tion is the circumstance that, apart from this occasional overlapping of the subject-matter, the broad distinctions between Brahmana Aranyaka and Upanishad are by no means always correctly observed; e.g., among the Aitareyins the matter of the Brahmana extends into the Aranyaka, while with the Taittiriyakas the close of the Brahmana and the beginning of the Aranyaka agree throughout, and the dividing line is entirely arbitrary. This state of things is to be explained probably only on the supposition that the entire teaching material of each S'akha formed originally a consecutive whole, and that this whole was first in the later times distinguished into Brahmana Aranyaka and Upanishad, on a principle which did not depend upon the character of the subject-matter alone, but which, though in general correspondence with it, was in fact imposed from without. Such a principle we seem to be able to recognise in the later order of the four As'rarnas, by virtue of which it became the duty of every Indian Brahman first as brahmac'dmn to spend a portion of his life with a Brahman teacher, then as grihastha to rear a family and to carry out the obligatory sacrifices, in order thereafter as vdnaprastha to withdraw into the solitude of the forest, and to devote himself to self- discipline and meditation, until finally in extreme old age. UPANISHADS OF THE THREE OLDER VEDAS 5 purified from all attachment to earth, homeless and with out possessions, free from all obligations, he wandered about as sannydsin (bhikshu, parivrdjaka), awaiting only his spirit's release into the supreme spirit. In the instruction communicated to him the brahmac'drin was put in posses sion of a rule of conduct for his entire future life. From the Brahmana he learnt how, as grihastha, he would have to carry out the ritual of sacrifice with the aid of the officiating priests ; the Aranyaka, as indeed is implied in the name, belonged to the period of life as vdnaprastha, during which for the most part meditation took the place of the sacrificial acts ; and finally the Upanishad taught theoretically that aloofness from the world which the sann ydsin was bound to realise in practice. Therefore it is said of him, that he should " live without the (liturgical) precepts of the Veda," but yet " recite the Aranyaka and the Upanishad of all the Vedas."1 And as ordinarily Aranyaka and Upanishad were blended together, so until quite late times, as we shall see, no strict line of demarcation was drawn in most instances between vdnaprastha and sannydsin. 3. TJie Upanisliads of the three older Vedas As the Brahmanas formed the ritual text-books of the Vedic S'akhas, so the Upanisliads attached to them were originally nothing more that the text-books of dogma, a fact which accounts especially for the identity in them all of the fundamental thought, which is developed at greater or less length and with the utmost variety. The earliest rise of the S'akhas or Vedic schools, on which this community of the ritual, and with it the philosophical tradition de pends, is to be sought in a time in which the contents of the Sariihita were already substantially fixed, and were trans mitted from teacher to pupil to be committed to memory.2 1 Aruneya-Up. 2. 2 Cp. Chand. 6. 7. 2. 6 THE PHILOSOPHY OF THE UPANISHADS On the other hand the necessary ritual allegorical and dogmatic explanations were communicated to the pupils extempore, and from these subsequently the oldest Indian prose took its rise. The result was that the common material of instruction, which in its essential features was already determined, received very various modifications, corresponding to the idiosyncrasy of the teacher, not only in regard to execution and mystical interpretation of the particular ceremonies, but also be cause one laid greater stress on the liturgical, another on the dogmatic teaching. Hence it is that the Upanishads of the individual schools differ so greatly in length. In the course of centuries the originally extempore instruction crystallised into fixed texts in prose, which were committed to memory verbatim by the pupil, while at the same time the divergences between the individual schools became wider. It is therefore quite credible that Indian writers should have been able to enumerate a considerable number of SYikhfis, in which each Veda was studied. But it is equally intelligible that of these many S'akhas the majority disappeared in the struggle for existence, and that for each Veda only a few prominent S'akhas with the Upanishads belonging to them have been preserved. We^nrust limit ourselves here for general guidance to a mere enumeration of the eleven extant Upanishads of the three older Vedas, with the remark, however, that in the case of several of these it is doubtful whether they are correctly attributed to the S'akha concerned. A further discussion of this point will be found in the Introductions prefixed to my translations of the sixty Upanishads. UPANIHHAD. I. Rigveda. Aitareya Upanishad. Aitareyins. Kaushitaki Upanishad. Kaushitakins. UPANISHADS OF THE ATHARVAVEDA 7 II. Samaveda. Chandogya Upanishad. Tandins. ^KenaT(TaIavakara) Upanishad. Jaiminiyas (Talavakaras). III. Yajurveda— (a) Black. Taittiriya Upanishad. ) „, . . A , Mahanarayana Upanishad. J Kathaka Upanishad. Kathas. S'vetas'vatara Upanishad. (wanting.) Maitrayaniya Upanishad Maitrayaniyas. (6) White. Brihadaranyaka Upanishad. ) Is'aUpanishad. J Vajasaneyms. 4. The Upanishads of the Atharvaveda The case is entirely different with the numerous Upa nishads which have found admission into the Atharva veda. It is true that several of them trace back their doctrine to S'aunaka or Pippalada, or even (as the Brahma-Up.) to both together ; and according to the tradition communicated by Narayana and Colebrooke, not only single treatises, but complete series of Upani shads were attributed to the S'aunakiyas or Pippaladis. But the contradictions of these accounts, as well as the circumstance that the most diverse Upanishads refer their doctrine to the alleged founders of the Atharvaveda S'akhas, S'aunaka and Pippalada, suggest the conjecture that we should see in this little more than an arbitrary attachment to well-known names of antiquity ; just as other Atharva-Upanishads trace back their doctrine to Yajnavalkhya, to Angiras or Atharvan, or even to Brahma Rudra and Prajapati. Moreover the names of the Atharva-Upanishads (apart from a few doubtful excep tions, as MdnddJcya, Jdbdla, Paihgala, Shavank) are no longer, as is the case with the Upanishads of the three older Vedas, formed on the model of the names of the S'akhas, but are derived partly from the contents and partly from any accidental circumstance. This proves that in the Atharva-Upanishads we must not 8 THE PHILOSOPHY OF THE UPANISHADS expect to find the dogmatic text-books of definite Vedic schools. Many indications (of which more will be said hereafter) point to the fact that the leading Ideas of the Upanishads, the doctrine, namely, of the sole reality of the Atman, of its evolution as the universe, its identity with the soul, and so forth, although they may have originated from Brahmans such as Yajnavalkhya, yet in the earliest times met with acceptance rather in Kshatriya circles1 than among Brahmans, engrossed as the latter were in the ritual. It was only later on that they were adopted by the Brahmans, and interwoven with the ritual on the lines of allegorical interpretation. Under these circumstances it is very probable that the atman doctrine, after it had been taken in hand by the S'akhas of the three older Vedas, was further prosecuted outside of these schools, and that consequently in course of time works were published, and have been partially at least preserved, which occupy a position as compared with the Upanishads of the Rig Sama and Yajurvedas precisely similar to that of the Samhita of the Athar- vaveda to their Samhitas. And as at an earlier date hymns of various kinds found admittance into this Samhita, which were partly of too late composition for the older Samhitas, and partly were despised by them ; so now again it was the Atharvaveda which opened its arms to the late born or rejected children of the spirit of atman research. The consequence of this generosity was that in course of time everything which appeared in the shape of an Upanishad, that is a mystical text, 1 As an illustration of the different relation of Brahmans and Kslmtriyas to the novel doctrine of the Atman, Brill. 3-4 may be referred to, where Yajnavalkhya, as exponent of this new doctrine, is met with jealousy and doubt on the side of the Brahmans, but by the king Janaka with enthusi astic assent. To this question we return later (infra, p. 17 ff.). UPANISHADS OF THE ATHARVAVEDA 9 whether it were the expression merely of the religious philosophical consciousness of a limited circle or even an individual thinker, was credited to the Atharvaveda, or by later collectors was included in it without further hesita tion. The regularity with which a given text reappears in the different collections forms, as far as we can see, the sole mark of its canonicity (if we may use the word in such a connection). Guided by this principle we have gathered together in our translation of the " Sixty Upani- shads " all those texts which seem to have met with general recognition. Eeferring then for further details to the Introduction there to the Atharva-Upanishads, we propose here, for the sake of a general survey, merely to enumerate the more important of these works according to the fivefold classification which we have made of them.1 I. PURE VEDANTA UPANISHADS. — These remain essenti ally faithful to the old Vedanta doctrine, without laying more definite stress than is already the case in the older Upanishads on its development into the Yoga, Sannyasa, and Vaishnavite or S'aivite symbolism :— Mundaka, Pras'na, Mdndukya (with the Karika) ; Garbha, Prdndgnihotra, Pinda; Atma, Sarvopanishatsdra, Gdruda. II. YOGA UPANISHADS. — These from the standpoint of the Vedanta treat predominantly and exclusively of the apprehension of the Atman through the Yoga by means of the morse of the syllable Om :— Bralimavidyd, Kshurikd, C'ulikd; Nddabindu, Bralimabindu, Amritabindu.Dliydna- bindu, Tejobindu ; s'ikhd, Yogatattva, Hariisa. III. SANNYASA UPANISHADS. — As a rule these are equally one-sided, and enjoin and describe the life 1 Following, in reality, Weber's example. io THE PHILOSOPHY OF THE UPANISHADS of the Sannyasin as the practical issue of Upanishad teaching :— Brahma, Sannydsa, Aruneya, Kanthas'ruti ; Paramahamsa, Jdbdla, As'rama. IV. S'IVA UPANISHADS. — These interpret the popularly worshipped S'iva (Is'ana, Mahes'vara, Mahadeva, etc.) as a personification of the Atmari : — Atharvasiras, Atharvas'ikhd, Nilarudra; Kdldgnirudra, Kaivalya. V. VISHNU UPANISHADS. - - These explain Vishnu (Narayana, Nrisirhha, etc.) similarly in the sense of the Upanishad teaching, and regard his various avataras as impersonations of the Atman : — Mahd, Ndrdyana, Atmabodha; Nrisimhapiirvatdpaniya, Nrisimhottaratdpantya ; Rdmaptirvatdpaniya, Rdmottaratdpaniya. 5. On the Meaning of the Word Upanishad According to S'ankara, the Upanishads were so named because they "destroy" inborn ignorance,1 or because they "conduct" to Brahman.2 Apart from these inter pretations, justifiable neither on grounds of philology nor of fact, the word Upanishad is usually explained by Indian writers by rahasyam (i.e. " secret," Anquetil's secretum tecjendum). Thus it is said, for example, in Nrisiihh. 8 four times in succession iti rahasyam, instead of the earlier usual form iti Upanishad (as is found e.g. at the close of Taitt. 2 and 3, Mahanar. 62. 63. 64). In older passages also, where mention is made of Upanishad texts, such expressions are used as guhyd' ddes'dh? paramam guli yam* vedaguhya-upanishatsu gildh am? guliyatamam* 1 S'ankara on Brih. p. 2. 4, Kuth. p. 73. 11. 2 Id. on Taitt. p. 9. 5, Mum), p. 2G1. 10. 3 Chfuid. 3. o. 2. 4 Kuth. 3. 17, S'vet. 6. 22. 5 S'vet. 5. G. 6 Maitr. G. 29. MEANING OF THE WORD UPANISHAD n The attempt to maintain secrecy with regard to abstruse and therefore easily misunderstood doctrines has numerous analogies even in the West. To the question why He speaks to them in parables Jesus answers, on SeSorai yvwvai ra /jLvarrfpia TT}? /3ao-tXe/a? r&v ovpai'&v, Se ov SeSorai.1 Pythagoras requires of his pupils (ricoTrr), mystical silence. A saying is preserved of Heracleitus, ra r^ yvwa-ea)*: {3d6ij Kpvrrreiv amari'T) ajaO^. Plato finds fault with the art of writing on the ground that it OVK erriararai \eyeiv o£? Set 76 real /j,ij.2 And Schopenhauer demands of his readers as a preliminary condition that they should have grappled with the diffi culties of Kant. The same feeling inspires the warning repeated again and again in the Upanishads, not to impart a certain doctrine to unworthy students. . Ait. Ar. 3. 2. 6. 9 : — " These combinations of letters (according to their secret meaning, their upanishad) the teacher shall not impart to anyone who is not his immediate pupil (antevdsin), who has not already lived for a year in his house, who does not himself intend to be a teacher." Chand. 3. 11. 5: — " Therefore only to his eldest son shall the father as Brahman communicate it (this doctrine), but to no one else, whoever he may be." Brih. 6. 3. 12:— "This (the mixed drink, mantha, and its ritual) shall be communicated to no one, except the son or the pupil." S'vet. 6. 22 : — " Give it (this supreme secret) to none who is not tranquil, who is not a son or at least a pupil." Mund. 3. 2. 11 : — " None may read this who has not observed his vow." Maitr. 6. 29 : — " This most mysterious secret shall be 13. 11. 2 Phaedr. 275, E. 12 THE PHILOSOPHY OF THE UPANISHADS imparted to none who is not a son or a pupil, and who has not yet attained tranquillity." Nrisiihh. 1. 3 : — " But if a woman or a S'udra learns the Savitri formula, the Lakshmi formula, the Pranava, one and all go downwards after death. Therefore let these never be communicated to such ! If anyone communicates these to them, they and the teacher alike go downwards after death." Ramap. 84 : — " Give it not (the diagram) to common men." The same explanation is to be given of the striking feature, which is constantly recurring in the Upanishads, that a teacher refuses to impart any instruction to a pupil who approaches him, until by persistence in his endeavour he has proved his worthiness to receive the instruction. The best known instance of this kind is Naciketas in the Kfithaka Upanishad, to whom the god of death vouchsafes the desired instruction on the nature of the soul and its fate only after the young man has steadily rejected all attempts to divert him from his wish.1 Indra deals in a similar way with Pratardana,2 Raikva with Janas'ruti,3 Satyakama with Upakosala,4 Pravahana with Aruni,5 Prajapati with Indra and Vairoc'aua,6 Yfijnavalkya with Janaka,7 S'akayanya with Brihadratha.8 From all this it follows that the universal tendency of antiquity, and of the circle which produced the Upanishads, was in the direction of keeping their contents secret from unfit persons, and that the Indian writers were practically justified in explaining the term upanishad by rahasyam, " secret." Less easy is it at first sight to understand how the word upanishad has 1 Kath. 1. 20 f. 2 Kaush. 3. 1. 3 Chand. 4. 2. 4 Chand. 4. 10. 2. 8 Chand. 5. 3. 7, Brih. 6. 2. 6. c Cband. 8. 8. 4. 7 Brih. 4. 3. 1 f. 8 Maitr. 1. 2. MEANING OF THE WORD UPANISHAD 13 come to signify " secret meaning, secret instruction, a secret." For upanishad, derived as a substantive from the root sad, to sit, can only denote a "sitting"; and as the preposition upa (near by) indicates, in contrast to parishad, samsad (assembly), a " confidential secret sitting," we must assume, even if actual proof is wanting, that this name for " secret-sitting" was used also in course of time to denote the purpose of this sitting, i.e. "secret instruction." Just as the German "college" has been transferred from the idea of " convention " to that of the subject-matter of instruction ; so that in such an expression as "to read, to hear, etc. a lecture" the original meaning of college (from colligere, to collect) is altogether forgotten, as in the case of the Upanishads the original conception of " sitting." Similar instances are quite common, as for example the faa-wal a/cpodo-eis of Aristotle or the Siarpipai of Epictetus no longer signify lectures, conversations, but definite written compositions. Another explanation of the word upanishad has been recently put forward by Oldenberg, according to which upanishad, precisely as updsand, would have originally meant " adoration," i.e. reverential meditation on the Brahman or Atman.1 The suggestion deserves attention, but is open to the following objections, (l) The words upa + as, "to sit before someone or something (in adora tion)," and upa -{-sad (upa + ni + sad does not occur in the Upanishads), " to seat oneself before someone (for the purpose of instruction)," are, according to prevailing usage, to be carefully distinguished from one another. Even if in the older texts the linguistic usage was not yet rigorously fixed, yet in the Upanishads (as a glance at Jacob's concordance proves), upa + as is always "to worship," never "to approach for instruction," and upa + sad always " to approach for instruction," never " to 1 Zeitschr. d. Deutsch. Morgenl. Gesellschaft, Bd. 50 (1896), p. 457 f. i4 THE PHILOSOPHY OF THE UPANISHADS worship " ; and the reason for forming the substantive upanishad not from upa + sad, but from the rarer upa + ni + sad, was perhaps merely that the substantive upasad had been already adopted as the name of a well-known ceremony preliminary to the Soma sacrifice. (2) Even if mention is frequently made of worship of Brahman or the atmau, especially under a definite symbol (as manas, prdna, etc.), yet, strictly speaking, the atman is not like the gods an object of worship, but an object of knowledge. Kcna 1. 4 f., — "that shouldest thou know as Brahman, not that which is there worshipped " (na idam yad idam updsate); Chanel. 8. 7. 1, — " the self (atman) . . .that ought man to search after, that endeavour to know " ; Brill. 2. 4. 5, — " the self, in truth, should be seen, heard, understood, and reflected upon, 0 Maitreyi," etc. The two passages of the Upanishads also, which Oldenberg cites in proof of worship offered to Brahman, tell in reality in the opposite direction. In Brih. 2. 1, Gargya declares his worship of this or that as Brahman, until finally the king breaks off the inquiry with the words, "with all that it is not yet known" (na etdvatd viditam bhavati). Then he imparts the teaching concerning the deep sleeper, and closes with the words, "his upanishad" (secret name, not worship) " is ' the reality of realities,' " i.e. the essence which is implied in all empirical existence. And if in Brih. 1. 4 the proposition is laid down that not the gods but the atman alone should be worshipped, by this is to be understood merely a polemic against the worship of the gods, not a demand to "worship" the atman as though it were only a god. This word is applicable, therefore, solely to the gods, and is used of the atman only by zeugma,1 and the proof of this is found 1 If this is disputed, then, to be consistent, from passages like Brih. 2. 4. 5, " the atman in truth should be seen and heard," etc., the conclusion must be drawn that the utman is visible and audible. MEANING OF THE WORD UPANISHAD 15 in what follows when it is said, — " He who worships another deity, and says ' He is one, and I am an other/ that man is not wise." 1 Without, however, such a conception of the atman as "He is one, and I am another," which is here interdicted, worship is altogether inconceivable, but not perhaps knowledge by immediate intuition (anubhava). (3) An attempt to apply the hypo thesis under consideration throughout to the existing facts would demonstrate its impossibility. Thus in Taitt. 1. 3 the secret meaning (upanishad) of the combination of letters (samhitd) is explained, and this being concluded various rewards are held out in prospect to him " who knows these great combinations as thus expounded " (ya evam etd mahdsamhitd vydkhydtd veda). Here merely a knowledge of the combination of the letters is required ; there is no mention of any worship in the entire paragraph. Or if we take the certainly ancient passage Kaush. 2. 1-2, where it is said of the beggar, who knows himself as the Self of all beings, — tasya upanishad ' na ydc'ed' iti, "his secret sign is not to beg"; it would be very difficult to say what suggestion of " worship " is found in phrases like these. If the passages collected in my index to the Upanishads under the word Upanishad are examined, it will be at once evident that, taken together, they involve the meaning, " secret sign, secret name, secret import, secret word, secret formula, secret instruction," and that therefore to all the meanings the note of secrecy is attached. Hence we may conclude that the explana tion offered by the Indians of the word upanishad as raliasyam, "secret," is correct, 1 Brih. 1. 4. 10. 1 6 THE PHILOSOPHY OF THE UPANISHADS II. BRIEF SUMMARY OF THE HISTORY OF THE UPANISHADS 1. The earliest Origin of the Upanishads The word Upanishad occurs with three distinct meanings as — (1) Secret word. (2) Secret text. (3) Secret import. ( 1 ) Certain mysterious words, expressions, and formulas, which are only intelligible to the initiated, are described as UpanisJiad. These contain either a secret rule for action and behaviour, as the na ydc'et of Kaush. 2. 1,2, quoted above, or secret information on the nature of Brahman. When, then, the latter is described as satyasya satyam1, or tad-vanam2 (the final goal of aspiration), there is added, " thou hast been taught the Upanishad." Of a similar nature are secret words like tajjaldn? " in him (all beings) are born, perish, and breathe," or neti neti.4 And when the worship of Brahman under such formulas is enjoined, it is not implied that upanishad signifies " worship," but only, as already pointed out, that medita tion on Brahman under these mysterious terms must take the place of the worship of the gods. (2) The extant texts themselves, as well as the older texts underlying them, are called Upanishads. Accord ingly in the Taittiriyaka school especially a section often ends with the words, — iti upanishad. (3) Very frequently it is not a word or a text, but the secret allegorical meaning of some ritual conception or practice, which is described as upanixhad-, e.g. in Chand. 1. 1. 10, — "for that which is executed with knowledge, * Brih. 2. 1. 20, 2. 3. G. 2 Ki>na 31 (4. G). 8 Chand. 3. 14. 1. 4 Brih. 2. 3. G, and often. EARLIEST ORIGIN OF THE UPANISHADS 17 with faith, with the upanishad (knowledge of the secret meaning of Udgitha as Om), that is more effective." The question suggests itself, which of these three significations is the original. We might decide for the third, and suppose that an allegorical interpretation was assigned to the ritual, and the Upanishad doctrine developed thence. This, however, apparently was not the case, and there is much to be said for the view that, as already observed above, the conceptions of the Upanishads, though they may have originated with the Brahmans, were fostered primarily among the Kshatriyas and not within Brahman circles, engrossed as these were with the ritual. The Upanishads have come down to us, like the rest of the texts of the three older Vedas, through the Brah mans. All the more striking is it, therefore, that the texts themselves frequently trace back some of their most important doctrines to kings, i.e. Kshatriyas. Thus, in the narrative of Chand. ^5. 11-24, five learned Brahmans request from Uddalaka Aruni instruction concerning the Atman Vais'vanara. Uddalaka distrusts his ability to explain everything to them, and all the six therefore betake themselves to the king As'vapati Kaikeya, and receive from him the true instruction, the defectiveness of their own knowledge having first been made clear. In Brih. 2. 1 (and the parallel passage, Kaush. 4), the far-famed Vedic scholar Gargya Balaki volunteers to expound the Brahman to King Ajatas'atru of Kasl, and propounds accordingly twelve (in Kaush. 16) erroneous explanations; whereupon to him, the Brahman, the king exhibits the Brahman as the atman under the figure of a deep sleeper, prefacing his exposition with the remark, " that is a reversal of the rule, for a Brahman to betake himself as a pupil to a Kshatriya in order to have the Brahman expounded to him ; now I proceed to instruct you." In this narrative, preserved by two different Vedic 1 8 THE PHILOSOPHY OF THE UPANISHADS schools, it is expressly declared that the knowledge of the Brahman as atman, the central doctrine of the entire Vedanta, is possessed by the king ; but, on the contrary, is not possessed by the Brahman "famed as a Vedic scholar." l In Chand. 1. 8-9, two Brahmans are instructed by the king Pravahana Jaivali concerning the dkds'a as the ultimate substratum of all things, of which they are ignorant. And although it is said in Chand. 1. 9. 3 that this instruction had been previously imparted by Atidhanvan to Udaras'andilya, yet the names allow of the conjecture that in this case also a Brahman received instruction from a Kshatriya. Similarly Chand. 7 contains the teaching given by Sanatkumara, the god of war, to the Brahman Narada. Here the former pronounces in adequate the comprehensive Vedic learning of the Brah man with the words : " all that you have studied is merely name."2 Finally the leading text of the doctrine of the soul's transmigration, which is extant in three different recensions,3 is propounded in the form of an instruction given to Aruni by the king Pravahana Jaivali.4 The king here says to the Brahman : — " Because, as you have told me, 0 Gautama, this doctrine has never up to the present time been in circulation among Brahmans, therefore in all the worlds the government has remained in the hands of the warrior caste."1 When we consider that the passages quoted discuss such subjects as the knowledge of Brahman as atman,6 the knowledge of this atman as the all-quickener,7 and the i Kaush., l.c. » Chand. 7. 1.3. 3 Chand. 5. 3-10, Brih. 5. 2, and with considerable variations Kaush. 1. 4 In Kaush., I.e., by C'itra Gangyfiyana. 'Chand. 5. 3. 7 ; in Brih. 6. 2. 8 the words are:— "As surely as I wish that you, like your ancestors, may remain well -disposed to us, so surely up to the present day this knowledge has never been in the possession of a Brahman." • Brih. 2. 1, Kaush. 4. T Chand. 5. 11 f. EARLIEST ORIGIN OF THE UPANISHADS 19 fate of the soul after death,1 that is, precisely the most important points of Upanishad teaching ; that not only is the king represented in them as endowed with wisdom, but is expressly contrasted with the Brahman who is ignorant or deluded ; and that these narratives are preserved to us by the Vedic Sakhas, and therefore by the Brahmans themselves ; we are forced to conclude, if not with absolute certainty, yet with a very high degree of probability, that as a matter of fact the doctrine of the atman, standing as it did in such sharp contrast to all the principles of the Vedic ritual, though the original concep tion may have been due to Brahmans, was taken up and cultivated primarily not in Brahman but in Kshatriya circles, and was first adopted by the former in later times. The fact, moreover, which is especially prominent in the last quoted passages, that the Brahmans during a long period had not attained to the possession of this knowledge, for which they nevertheless display great eagerness, is most simply explained on the supposition that this teach ing with regard to the atman was studiously withheld from them ; that it was transmitted in a narrow circle among the Kshatriyas to the exclusion of the Brahmans ; that, in a word, it was upanishad. The allegorical method of interpreting the ritual in the light of the atman doctrine, though it may have been already practised among the Kshatriya circles, was probably undertaken on a larger scale after the adoption of the new doctrine by the Brahmans. It would follow that the third of the above-mentioned meanings of the word upanishad as " secret import " (of some ritual conception) is probably in the first instance secondary. If we ask further, which of the two other meanings, (1) secret word, (2) secret text, is the more primitive, it would seem that a transition from the second to the first is with difficulty intelligible, 1 Chand. 5. 3 f., Brih. 6. 2. 20 THE PHILOSOPHY OF THE UPANISHxVDS but that the first passes into the second by a natural and readily comprehended change. We may therefore assume that the doctrine of the atman as the first principle of the universe, the gradual rise of which we have traced through the hymns of the Rigveda and Atharvaveda, was fostered and pro gressively developed by the Kshatriyas in opposition to the principles of the Brahmanical ritual ; whence the new knowledge was expressed in brief words or formulas, intel ligible only to the initiated, such as tadvanam, tajjaldn, satyasya sat yam, samyadvdma, vdmani, bhdmant, etc. A formula of this kind was then called an upanisliad, inasmuch as the condition of its communication and ex planation was the absence of publicity. Such formulas were naturally accompanied by oral explanations, which also were kept secret, and from these were gradually developed the earliest texts that bore the name of Upanishad. The manner in which the formulas tad vai tad l or vi-ram 2 are discussed may serve as examples of such secret words accompanied by secret explanation.3 In these and similar ways the secret doctrines, i.e. the vidyds, arose, of which mention is so frequently made in the Upanishads. Their authors or exclusive possessors were renowned in the land. Pilgrims sought them, pupils served them for many years,4 and rich gifts were offered to them 5 in order thereby to gain the communication of the i Brih. 5.4. » Brih. 5. 12. 8 The explanations given of these secret words are not always in agreement. The definition of Brahman as pArtiam apravarti is approved in Chand. 3. 12. 7, but in Brili. 2. 1. 5 (Kaush. 4. 8) is regarded, on the contrary, as inadmissible. Of still greater interest is the case of the Upanishad Brih. 16. 3, amritam satyena c'hannam, understood by others as anritam satyena c'hannam ; so also Brih. 5. 5. 1 (anritam ubhayatah satyena parigrihttam), which again is otherwise explained in Chand. 8. 3. 5. Similarly the saying of the ancient rishis, pdhktam idam sarvam, is differently construed in Brih. 1. 4. 17 and Taitt. 1. 7. « Chand. 4. 10. 2, 5 Chand. 4. 2. 1. EARLIEST ORIGIN OF THE UPANISHADS 21 vidyd. In the case of some of these vidyds the name of the author is preserved. Several of them, in fact, are equipped with a formal genealogy, which recounts the original author and his successors, and usually closes with the injunction to communicate the doctrine only to a son or trusted pupil. A suitable field, however, for the successful development of these doctrines was first opened up when they passed from the Kshatriya circles, where they had originally found a home, by ways that a few illustrations have already taught us to recognise, into the possession of the Brahmans, whose system of scholastic traditions was firmly established. The latter eagerly adopted the atman doctrine, although it was fundamentally opposed to the Vedic cult of the gods and the Brahmanical system of ritual, combined it by the help of allegorical interpreta tion with the ritualistic tradition, and attached it to the curriculum of their schools. The Upanishads became the Vedanta. Soon also the Brahmans laid claim to the new teaching as their exclusive privilege. They were able to point to princes and leaders, as Janaka, Janas'ruti, etc., who were said to have gone for instruction to Brahmans. Authorities on the ritual like S'andilya and Yajnavalkhya were trans formed into originators and upholders of the ideas of the Upanishads, and the atman doctrine was made to pre suppose the tradition of the Veda : — " Only he who knows the Veda comprehends the great omnipresent Atman," as it is said in a passage of the Brahmanas.1 After the Upanishad ideas had been adopted by the S'akhas, and had been made a part of their Vedic system of instruction, they passed through a varied expansion and development under the hands of the Vedic teachers. To begin with they were brought into accord with the ritua] 1 Tailt. Br. 3. 12. 9. 7. 22 THE PHILOSOPHY OF THE UPANISHADS tradition by interpreting the latter (in the Ar.anyakas) in the spirit of the atman doctrine ; and thus the adherents of the Rigveda brought it into connection with the uJctham (hymn), those of the Sanaa veda with the sdman, and those of the Yajurveda with the sacrifice, especially the horse-sacrifice as being its highest form. The new doctrine, however, was further developed in a manner which altogether transcended the traditional cult, with which, indeed, it often found itself in open contradiction. In regard to this an active communication and exchange o o must have existed between the different schools. Defini tions which by the one were highly regarded failed to meet with acceptance in another. Teachers who in the one S'akha exercised supreme authority are found in an other in a subordinate position (Arum), or are altogether unknown (Yajnavalkhya). Texts appear with slight variations in the different Vedic schools, whether borrowed directly or going back on either side to a common original. Other texts are met with side by side in one arid the same S'akha in numerous recensions, often very similar, often widely divergent from one another. This rich mental life, the details of which can scarcely be further reproduced, may not improbably have lasted for centuries ; and the fundamental thought of the doctrine of the atman have attained an ever complete! development by means of the reflection of individual thinkers in familiar intercourse before a chosen circle of pupils, and probably also by public discussions at royal courts. The oldest Upanishads pre served to us are to be regarded as the final result of this mental process. 2. The extant Upanishads Owing to the manner in which the Upanishads have arisen from the activity of the different Vedic schools and their intercourse one with another, we are unable to lay THE EXTANT UPANISHADS 23 down any. precise chronological order of succession among them. All the principal Upanishads contain earlier and later elements side by side, and therefore the age of each separate piece must be determined by itself as far as this is possible from the degree of development of the thoughts which find expression in it. Here, where we still treat of the Upanishads as a whole, we can only attempt a rough and approximate determination of the period to which in general an Upanishad belongs. We distinguish first four successive periods of time, to which the Upanishads as a whole may be assigned. I. THE ANCIENT PROSE UPANISHADS.— Brihadaranyaka and Chandogya. Taittiriya. Aitareya. Kaushitaki. Kena. The last-named stands on the border-line. These are collectively the Vedanta texts of the actually existing S'akhas, and in their earlier parts are usually closely interwoven with Brahmanas and Aranyakas, of which they form the continuation, and whose ritualistic conceptions are interpreted by them in various allegorical ways. It is only the later, and as we may suppose younger texts which emancipate themselves from the ritual. The language is still almost entirely the ancient prose of the Brahmanas, somewhat ponderous stilted and awkward, but not without natural charm. The order adopted above is in general chronological. The Brihadaranyaka and Chandogya are not only the richest in contents, but also the oldest of the extant Upanishads. As compared also with one another, the Brihaddranyaka, as we shall often see, shows almost without exception greater originality in the grouping of the texts. On the other hand the literary outlook of Chand. 7. 1. 4 (7. 2. 1, 7. 7. 1) is materially 24 THE PHILOSOPHY OF THE UPANISHADS broader than that of Brih. 2. 4. 10 (4. 1. 2, 4. 5. 11). Taittirtya in its essential part is still later than Chan- dogya ; cp. Chand. 6. 2 (three elements) and Taitt. 2. 1 (five elements). Aitareya is later than Chdndogya (in Chand. 6. 3. 1 there are three kinds of organic beings, in Ait. 3. 3 four), and than Taittirtya (cp. Taitt. 2. 6, "after that he had created it he entered into it," with the more elaborate description Ait. 1. 3. 12). KaushUaki,&iinlly, is later than all those named ; for Kaush. 1 is less original than Chand. 5. 3 f., Brili. 6. 2, and Kaush. 3 must be later than Ait. 3. 3, Kaush. 4 than Brih. 2. 1. Kena stands on the border-line of this period, and by virtue of its first metrical portion already belongs to the succeeding epoch. II. THE METRICAL UPANISHADS. — The transition is made by Kena 1-13 and the verses Brih. 4. 4, 8-21, undoubtedly a later addition. There follow Kathaka Isi S'vetas'vatara. Mundaka. Mahanarayana. The last-named makes use of Mundaka, and Mundaka appears to use Svetds'vatara. Is'd seems on the whole to be less fully developed than Svetds'vatara, and to be freer from sectarian bias ; but in numerous instances it is found to be dependent oil Kdthaka.1 That Svetds'vatara is later than Kdthaka is not open to doubt; on the contrary, it is very probable, on the evidence of several passages,2 that Kdthaka was directly employed in the com position of Svetds'vatara. The difference between this period and the preceding is very great. The connection with the S akhas appears 1 Cp. especially IsTi 8 with Kath. 5. 13. 2 Collected in Deussen, Upan., p. ^89. THE EXTANT UPANISHADS 25 sometimes doubtful, sometimes artificial, and in any case is loose. Allegories framed after the mariner of the Aranyakas are wanting. The thought of the Upanishads is no longer apprehended as in course of development, but appears everywhere to have been taken over in its entirety. Individual verses and characteristic phrases con stantly recur. The phraseology is already formed. And the language is almost throughout metrical. III. THE LATER PROSE UPANISHADS. — Pras'na. Maitrayaniya. Mandukya. In this third period the composition returns again to prose, but a prose which is markedly different from the archaic language of the ancient Upanishads, although it does also take on, especially in theMaitrdijamija, an archaic colouring. The style suggests that of the later Sanskrit prose ; it is complex, involved, and delights in repetitions. The dependence of the thought on that of the earlier Upani shads is made manifest by numerous quotations and adap tations. That Pras'na is later than Mundaka is proved by the fact that the latter is quoted in Pr. 3. 5 ; it is older, however, than Maitrayaniya, for it is itself quoted in Maitr. 6. 5. The position of Mdndakya is difficult to determine, owing to its brevity ; yet the theory concerning Om in Hand. 3 seems to be more advanced than that of Maitr. 6. 4. The greater number of the Upanishads hitherto mentioned have found admission, sometimes with very doubtful right, to a place in the three older Vedas. Only three of them— namely, Mundaka, Pras'na, and Mandukya— appear to have belonged from the beginning to the Atharvaveda, the two first-named certainly as the original legitimate Upanishads of this fourth Veda. These two are ascribed to S'aunaka and Pippalada, the founders of the STikhas of the Atharva- 26 THE PHILOSOPHY OF THE UPANISHADS veda. The later collections of Atharva Upanishads begin as a rule with the Mundaka and Pras'na, and these two alone can he proved to have been known to and employed by Badarayana and S'ankara. IV. THE LATER ATHARVA UPANISHADS. — Later theo logical treatises retain still the form of Upanishads as a convenient method of literary composition that carries with it a degree of sanctity ; while the thought concerns itself partly with the continuous development of older themes, or refrains from deviating from the beaten tracks (Garbha, Prdndgnihotra, Pinda, Atma, Sarvopanishat- sdra, Gdruda), partly turns its attention to the glorifi cation of the Yoga (Brahmavidyd, Kshurikd, Cdlikd, Xddabindu, Brahmabindu, Amritabindu, Dhydnabindu, Tcjobindu, Yogasikhd, Yogatattva, Hamsa), or of the Sanuyasa (Brahma, Sannydsa, Aruneya, Kanthas'ruti, Paramahamsa, Jdbdla, As'rama). The difference between the two tendencies shows itself also in the fact that almost without exception the Yoga Upanishads are com posed in verse, those of the Sannydsa in prose with occasional verses inserted. A further class of Upanishads is devoted to the worship of S'iva (Atharvas'iras, Atharva- s-ikhd, Xilarudra, Kdldgnirudra, Kaivalya), or of Vishnu (Mahd, Ndrdyana, Atmabodha, Nrisimhatd- paniya, Rdmatdpantya, and endeavours to interpret these in the light of the atman doctrine. They are composed for the most part in prose with an inter mixture of verse. All of these Upanishads were received into the Atharvaveda, but met with no recognition from the leading theologians of the Vedanta. 3. The, Upanishads in Bddardyana and Sankara The earliest traces of a collection of Upanishads are found within the books themselves. Thus the mention in S'vet 5. 6 of "the Upanishads that form the mystical BADARAYANA AND S'ANKARA 27 portion of the Veda" (veda-guhya-upanishadah), and also the passage S'vet. 6. 22, "in former times in the Vedanta was tlie deepest mystery revealed," seem to look back to the older Upanishads as a self-contained whole which already claimed a certain antiquity. A similar inference may be drawn from a thrice recurring verse1 which speaks of ascetics (yatis) who have " grasped the meaning of the Vedanta doctrine." Still more clearly do the Upanishads appear as a complete whole when, in Maitr. 2. 3, the doctrine concerning Brahman is described as " the doctrine of all the Upanishads " (sarva-upanishad- vidya). That in so late works as the Sarva-upanishad-sara or the Muktika Upanishad the Upanishads are assumed to be a whole is therefore of no further importance. It was undoubtedly on the foundation of older and earlier works that Badarayana formally undertook an epitome of Upanishad doctrine in the Brahmastitras, the foundation of the later Vedanta. He shows that Brahman is the first principle of the world, samanvaydt, "from the agreement" of the Upanishad texts,2 and proclaims the fundamental proposition " that all the texts of the Vedanta deserve credence " (sarva-veddnta- pratyayam}? Which Upanishads, however, were recog nised by him as canonical cannot be ascertained from the sutras themselves owing to their brevity, but only from S'ankara's commentary, and the decision therefore remains in many instances doubtful, since we do not know how far S'ankara followed a reliable tradition. Only in the first adhydya is it possible to determine with greater certainty the Upanishad texts which Badarayana had in his mind, where he undertakes to establish the teaching concerningBrahman in twenty-eight Adhikaranas (sections) based on as many passages of the Upanishads. 1 Mund. 3. 2. 6, Malumur. 10. 22, Kniv. 3. 2 1. 1. 4. 3 3. 3> L 28 THE PHILOSOPHY OF THE UPANISHADS Here, as in his entire work, the number four plays a decisive part in the arrangement of the material. Of the twenty-eight fundamental passages, twelve are taken from the Chandogya, four from the Brihadfiranyaka, four from the Kathaka, four from the Taittiriya and Kaushitaki (two from each), and four from the Atharva Upanishads, namely, three from the Mundaka and one from the Pras'na. The following scheme1 shows that the order of the passages, as they are found within each of the Upanishads which he employs, is strictly observed, while in other respects the passages appear interwoven in a manner for which we seem to be able to find a reason here and there in the close connection of the subject-matter. (1) 1. 1. 12-19. Taitt. 2. 5. (2) 20-21. Chand. 1. G. G. (3) 22. Chand. 1. 9. 1. (4) 23. Chand. 1. 11. 5. (5) 24-27. Chand. 3. 13. 7. (G) 28-31. Kaush. 3. 2. (7) 1. 2. 1-8. Cliand. 3. 14. 1. (8) 9-10. Kath. 2. 25. 0) 11-12. K:\tll. 3. 1. (10) 13-17. Chand. 4. 15. 1. (11) 18-20. Brih. 3. 7. 3. (12) 21-23. Mund. 1. 1. G. (13) 21-32. Chand. 5. 11-24. (14) 1. 3. 1-7. Mund. 2. 2. 5. (15) 8-9. Chand. *7 23. (1C) 10-12. Brill. 3. 8. 8. (17) 13. Pras'na, 5. 5. (18) 14-18. Chand. 8. 1. 1. (19) 1 9-2 1 . Chand. 8. 1 2. 3. (20) 22-23. Mund. 2. 2. 10. (21) 24-25. Kath. 4. 12. (22) 39. Kath. G. 1. (23) 40. Chand. 8. 12. 3. (24) 41. Chand. 8. 14. (25) 42-43. Brih. 4. 3. 7. (2G) 1. 4. 14-15. Taitt. 2. G. (27) 10-18. Kaush. 4. 19. (28) 19-22. Brih. 4. 5. 6. From Deussen, Systtnn. des Veddnta, p. 130. BADARlYANA AND S'ANKARA 29 The striking preference for the Chandogya suggests that an earlier work due to the school of this Upanishad was already in the hands of Badarayuna, into which he or one of his predecessors worked sixteen extracts of importance derived from another Sakha, being guided further by the principle that the original order of the extracts should be maintained. Besides the Upanishads named, Badarayana may with some confidence be shown to have used the S'vetas'vatara,1 Aitareya,2 and perhaps Jabala.3 With regard, however, to the formula of impre cation quoted in Sut. 3. 3. 25, which according to S'ankara should find a place " at the beginning of an Upanishad of the Atharvanikas," and which is nowhere known to exist, I would now suggest (since throughout their works Badarayana and S'ankara make use only of the Mundaka and Pras'na from the Atharva Upanishads, consequently recognise none but these, and since they appear to recog nise the authority of the Upanishad that follows the imprecation formula), that the suspected formula may once have stood at the beginning of one of these two, perhaps of the Mundaka Upanishad ; somewhat after the manner in which the S'anti formulas precede the Upani shads in some manuscripts, and in others are wanting. To the Brahmasutras of Badarayana is attached the great commentary of S'ankara (circa 800 A.D.), to whom are ascribed, besides other works, the commentaries on the Brihaddranyaka, Clidndogya, Taittiriya, Aitareya, S'vetas'vatara, Is'd, Kena, Katlia, Pras'na, Mundaka and Mdndukya, which are edited in the BiU. Ind., vols. ii., iii., vii., viii. Commentaries therefore of S'ankara are missing on the KausMtaki, which was first elucidated by S'ankardnanda (a teacher, according to Hall, Index, p. 98. 123, of Madhava, who flourished 1350 A.D.), and on the Maitrdyaniya, which Rdmatirtha expounded. 1 SAt- 1- 4. 8-10. 2 S6tt 3 3> 16_ia 3 SAt 30 THE PHILOSOPHY OF THE UPANISHADS The commentaries, however, on the eleven Upanishads named are to be attributed in part probably not to S'ankara himself, but merely to his school, since the explanations given in the Upanishad commentaries often fail to agree with those in the commentary on the sutras. The commentary on the Mdndtikya which is extant under the name of S'aukara treats this and Gaudapada's Karika as one, and seems to regard the whole as in no sense an Upanishad (p. 330 : veddnta-artha-sdra-san- graha-bhfitam idam prakarana-c'atuslitayam ' om iti etad aksharam' ityddi drabhyate)\ and with this would agree the fact that the Mdnddkya is not quoted either in the Brahmasutras or in S'ankara's commentary on them, wliile two verses from the Karika of Gaudapada1 are cited by S'ankara2 with the words, atra uktam veddnta-artha-sampraddyavidbliir dc'dryaih. In his commentary on the Brahmasutras only the following fourteen Upanishads can be shown to have been quoted by S'ankara (the figures attached indicate the number of quotations), — Chandogya 809, Brihadaranyaka 565, Taittiriya 142, Munrlaka 129, Kathaka 103, Kaushitaki 88, S'vctas'vatara ^53, Pras'na 38, Aitareya 22, Jabiila 13, ^Mahanarayana 9, Is'a 8, Paingi 6, Kena 5. Although S'ankara regards the texts of the Vedanta which he recognises as a uniform and consistent canon of truth,3 yet he seems still to have had in his hands no 1 3. 15 and 1. 10. 2 P. 375. 3, 433. 1. 3 We may compare his exposition on sutra 3. 3. 1, p. 843: — "How then can the question arise, whether the doctrines concerning the atman are different or not different ; for we cannot suppose the aim of the Vedanta is to teach a plurality of Brahmans, like the existing plurality of phenomena, since Brahman is one and immutable. So it is not possible that concerning the immutable Brahman various doctrines should exist ; for to suppose that the actual fact is one thing, and the knowledge of it another, is necessarily a mistake. And even supposing that in the different Vedfinta texts various doctrines were taught concerning the immutable Brahman, only one of these could be true ; the remainder on the other hand would be false, and the con- BlDARlYANA AND S'ANKARA 31 collection of Upanishads, since he looks upon the greater number of them as still forming the concluding chapters of their respective Brahmanas, to which therefore he is accustomed to refer at the commencement of the Upanishad commentary. Thus in the introduction to the commentary on the Kena1 he quotes its beginning as "the beginning of the ninth adhyaya ;2 before it works have been thoroughly discussed ; the acts of adora tion also of the prana which serves as the foundation of all works were taught ; and further those also which relate to the Sam an that forms a branch of the works. Next followed the consideration of the Gayatra-saman, and finally the list of teachers. All the above belongs still to works," etc. On Chdndogya, p. 2 : — " The entire ritual has been rehearsed, as also the knowledge of Prana- Agni, etc., as divine." On Taittiriya, p. 2: — "The appointed works which serve to atone for trangressions that have been committed, as also the works desirable for those who covet a definite reward, have been rehearsed in the pre ceding parts of the book (purvasmin granthe)." On Brihaddranyaka, p. 4 : " The connection of this (Upani shad) with the sphere of works is as follows," etc. On fs'd, p. 1 : — "The mantras is'd vdsyam, etc., do not apply (as we should expect) to works, but reveal the nature of the atmari who is independent of works." On Aitareya, p. 143 : — "The works together with the knowledge rela tive to the lower Brahman are remitted," etc. As may be inferred from the comments quoted, all these Upanishads appear to have been still regarded by S'ankara as the concluding portions of their respective sequence would be loss of confidence in the Vedanta. (This, however, in Sankara's eyes would be an arraycoyi} els TO aSvvarov}. It is therefore inconceivable that in the individual texts of the Vedanta a difference of doctrine on the subject of Brahman should find a place." 1 Bill Ind., p. 28. 2 In the recension published by Ortel it belongs to the fourth adhyaya. 32 THE PHILOSOPHY OF THE UPANISHADS Br&hmanas. On the other hand a similar connection with the part of a preceding work is wanting in the case of the commentaries on Kathaka and S'vetas'vatara. So also with Mundaka and Pras'na, which are treated by S'ankara as one. In the introduction to Pras'na, p. 160. 2, he remarks: — "in order to examine further the subjects taught in the mantras (of the Mundaka Upanishad, as it is rightly glossed), this Brahmana (the Pras'na Upanishad) is undertaken." Since, however, the Mundaka and Pras'na exhibit no relationship at all, and since they are attached further to different S akhfis of the Atharvaveda (those of S'aunaka and Pippalada respectively), this unity under which S'ankara treats of them is probably to be explained merely from the fact that as early as his time they were linked together as the first beginning and foundation of a collection of Atharva Upanishads. At that time probably the collection consisted only of these two, for otherwise it is hardly likely that the others would have been ignored by S'ankara so completely as was in fact the case. It is true also that the annotator Anandajndna remarks at the beginning of S'ankara's commentary on the Mdndukya : " Beginning with the Brahma Upanishad (he intends probably the Brahma-vidya Upanishad) and the Garbha Upanishad, there are extant besides many Upanishads of the Atharvaveda. Since, however, they are not em ployed in the S'arirakam (the Brahmasutras of Badarayana), lie (S'ankara) does not expound them." But the reason assigned is perhaps not conclusive ; for which Upanishads are found in the S'arirakam, and which not, could only be determined by tradition or from S'ankara himself. It must therefore have been tradition or S'ankara himself that excluded other Upanishads from the Canon, whether because they were yet unknown, or because they were not yet recognised as Upanishads. And thus in fact S'ankara describes the Mandukya, upon which nevertheless, together IMPORTANT COLLECTIONS OF UPANISHADS 33 with Gaudapada's Karika, he had himself commented, not as an Upanishad, but as " a literary composition contain ing the essence of the Vedanta (veddnta-artha-sdra- sahgraha-bhtltam prakaranam ). 4. The most important Collections of Upanishads The further history of the Upanishad tradition is for a time shrouded in darkness, and only conjecturally are we able from the existing collections of Upanishads to draw some conclusions as to their origin. These collections or lists fall from the outset into two classes, in so far as they either contain the Upanishads in their entirety, or limit themselves (at least as far as the original design is concerned) to the Upanishads of the Atharvaveda. Of the former class is the Canon of the Muktika and the Oupnek'hat, of the latter that of Colebrooke and Narayana. Since the Upanishads of the three older Vedas con tinued to live in the tradition of the S akhas, as long as these survived the secure transmission of the Upanishads concerned was assured. It was otherwise with the Atharvaveda, which was not employed at the sacrifice, and in consequence had no such firmly established tradition of the schools as the text of the three older Vedas upon which to rely for its preservation. This is shown not only by the indifference from which its Samhita has suffered, but also by the freedom with which it ad mitted new compositions. The latter would assuredly have been impossible as long as the tradition was under the protection of regular Vedic schools, maintaining them selves from generation to generation according to the rules of their guild. Hence is to be explained the exten sive irruption of newly composed Upanishads into the Atharvaveda. As early as Sankara we find the Mundaka and Pras'na united together (sup. p. 32), and on these as 3 34 THE PHILOSOPHY OF THE UPANISHADS foundation a collection of Atharva Upanishads appears to have been gradually built up, which eventually comprised 34 pieces from Mundaka to Nrisirhhatdpantya, and included also some whose claim to the name of Upani shads had never been previously recognised; just as in the judgement of S'ankara the Kfirika of Gaudapada on the °Mandukya Upanishad, and indeed this treatise itself (sup. pp. 30, 33), had no claim to the position of an Upanishad. These 34 primary Upanishads of Colebrooke's list were later extended to 52 by the addition not only of a number of recent compositions, but most remarkably by the side of and among them of seven of the recognised texts of the older Vedas, viz.— 35-36 KtUhaka, 37 Kena, 39-40 Brihannarayana ( = Taitt. Ar. x.), 44 AnandavalU ( = Taitt. Up. 2), and 45 Bhriguvalll ( = Taitt. Up. 3). In this manner the collection of 52 Upanishads first made known by Colcbrooke originated, the strange combination of which we attempted to explain l on the hypothesis that at the time and in the region where this collection was finally put together the three older Vedas were cultivated only in the Sakhas of the Aitareyins, Tfmdins (to which the Chandogya Upanishad belongs), and Vajasaneyins. Accordingly the Upanishads of the remaining Sakhas (with the exception of the Kaushttaki, S'vetds'vatara, and Maitrdyanttja, which were perhaps already lost or not recognised) were inserted in the existing collection of Atharva Upanishads with a view to their safe pre servation.2 The collection of Nfirayana is in exact agreement wit that of Colebrooke, apart from a few variations in the i Deussen, Upan., p. 537. . . * An apparently older list has been preserved in the Atharva-paris. 6 vdc'drambhanam vikdro, ndmadheyam. 44 THE PHILOSOPHY OF THE UPANISHADS which meii regard as real is mere name.1 Later passages employ language that is based on these conceptions, " nor is this even a plurality,"2 and the verses preserved in Brih. 4. 4. 19 :3- In the spirit should this be perceived, Here there is no plurality anywhere. From death to death again he rushes blindly Who fancies that he here sees difference. Apt and striking also is the remark of a later Upanishad 4 that no proof of plurality can even be offered, " for no proof is possible of the existence of a duality, and only the timeless atman admits of proof," (i.e. we are incapable of knowing anything outside of our own con sciousness, which under all circumstances forms a unity). It is clear from the foregoing: — (1) That the view which later was most explicitly set forth in the doctrine of mdyd is so far from being strange to the oldest Upanishads that it is assumed in and with their funda mental doctrine of the sole reality of the atman, and forms its necessary complement ; and (2) that this funda mental doctrine of the Upanishads is seen to be in mar vellous agreement with the philosophies of Parmenides and Plato, and of Kant and Schopenhauer. So fully indeed is this true, that all three, originating from different epochs and countries, and with modes of thought entirely independent, mutually complete, elucidate, and confirm one another. Let this then suffice for the philosophical significance of the Upanishads. 3. The Conception of the Upanishads in its Relation to Religion The thought referred to, common to India, Plato, and Kant, that the entire universe is only appearance and not 1 ro) Trai/r' 'oi/o/i' ttrrat, oo-